Graph of Graphs

Modeling relations and networks by graphs gives access to an immensely useful set tool. Graph theory provides the basis for powerful algorithms and their analysis with a wide range of applications. Relations between graphs can be viewed themselves as graphs as well and thus their analysis can also benefit from these tools. In this talk we investigate the graph of graphs, where the set of nodes comprises all labelled graphs and two nodes are adjacent if the edge sets of the corresponding graphs differ by exactly one edge. This graph of graph exhibits many intriguing properties and provides a different view on dynamic networks which can be seen as paths on the graph of graphs. Assigning costs to the edges of the graph of graphs allows us to define similarity measures with applications for temporal graph analysis. I demonstrate how these new tools can be exploited on real-world networks and discuss insights gained from them.

Yvonne-Anne Pignolet is a Principal Scientist at ABB Corporate Research, Baden, Switzerland since 2011, working on theoretical and practical aspects distributed computing, networking, and their application to automation systems. She holds a PhD (2009) and a Master's degree (2006) in Computer Science from ETH Zurich, Switzerland. During her Postdoc at IBM Research Zurich Laboratory, Switzerland, she was involved in projects on wireless sensor networks, and at Ben Gurion University of the Negev, Israel, she worked on models and analysis for complex networks. Since 2013, she has been teaching a course on Industrial Automation at EPFL, Switzerland.